Choosing a Primary Metric and Guardrails for YouTube's Home Feed

Quick Overview

A product manager metrics question about the YouTube homepage, the recommended feed people see when they open the app. It tests defining the page's purpose, choosing a primary metric for valuable viewing, and building supporting and guardrail metrics for viewers, creators, advertisers and page performance.

Choosing a Primary Metric and Guardrails for YouTube's Home Feed

Company: American Express

Role: Product Manager

Category: Analytics & Experimentation

Difficulty: medium

Interview Round: Other

You are the product manager responsible for YouTube's homepage: the personalized feed of recommended videos people see when they open YouTube. Which metrics would you use to judge whether the homepage is doing its job? Identify the metric you would treat as primary, and explain the supporting and guardrail metrics around it. ```hint Clicks are not the goal Start from what the homepage is for: helping a viewer find something worth watching. A high click rate on its own can reward thumbnails that disappoint once the video starts. ``` ### Clarifying Questions - Which surfaces are in scope: the phone app, desktop, the TV app, or all of them? Are signed-out viewers included? - Are these metrics for tracking the homepage's overall health, or for evaluating a specific change such as a new ranking model or layout? - Should the metrics cover only viewers, or also the creators and advertisers who depend on homepage exposure? - Do the Shorts shelves and ads that appear on the homepage fall within scope? ### What a Strong Answer Covers - The homepage's purpose and the viewer journey it supports, from opening YouTube to finishing a video and coming back. - A primary metric that captures valuable viewing started from the homepage, defined precisely. - Supporting metrics across reach, discovery, viewing quality and retention. - Guardrails for creators, advertisers, content responsibility and page performance. - How the metrics are segmented by device and viewer type, and how they are used in experiments. ### Follow-up Questions - After a ranking change, homepage click-through rate rises but watch time per click falls. How do you read that? - How would your metrics differ between the TV app and the phone app? - How would you detect that the homepage has become too narrow, showing people the same kind of content again and again?

Overview: A product manager metrics question about the YouTube homepage, the recommended feed people see when they open the app. It tests defining the page's purpose, choosing a primary metric for valuable viewing, and building supporting and guardrail metrics for viewers, creators, advertisers and page performance.

American Express logo
American Express
Nov 6, 2025
mediumProduct ManagerOtherAnalytics & Experimentation
0
0

You are the product manager responsible for YouTube's homepage: the personalized feed of recommended videos people see when they open YouTube. Which metrics would you use to judge whether the homepage is doing its job? Identify the metric you would treat as primary, and explain the supporting and guardrail metrics around it.

Clarifying Questions Guidance

  • Which surfaces are in scope: the phone app, desktop, the TV app, or all of them? Are signed-out viewers included?
  • Are these metrics for tracking the homepage's overall health, or for evaluating a specific change such as a new ranking model or layout?
  • Should the metrics cover only viewers, or also the creators and advertisers who depend on homepage exposure?
  • Do the Shorts shelves and ads that appear on the homepage fall within scope?

What a Strong Answer Covers Guidance

  • The homepage's purpose and the viewer journey it supports, from opening YouTube to finishing a video and coming back.
  • A primary metric that captures valuable viewing started from the homepage, defined precisely.
  • Supporting metrics across reach, discovery, viewing quality and retention.
  • Guardrails for creators, advertisers, content responsibility and page performance.
  • How the metrics are segmented by device and viewer type, and how they are used in experiments.

Follow-up Questions Guidance

  • After a ranking change, homepage click-through rate rises but watch time per click falls. How do you read that?
  • How would your metrics differ between the TV app and the phone app?
  • How would you detect that the homepage has become too narrow, showing people the same kind of content again and again?